A contingency model of the adoption of the multidivisional ...
Transcript of A contingency model of the adoption of the multidivisional ...
Digitized by the Internet Archive
in 2011 with funding from
University of Illinois Urbana-Champaign
http://www.archive.org/details/contingencymodel1577maho
BEBRFACULTY WORKINGPAPER NO. 89-1577
A Contingency Modelof the Adoption of the
Multidivisional Organization
a*.., * u \WUnrvorsuy o! Illinois
of Urtoana-Champalgn
Joseph T. Mahoney
Jte Library of the
AUG T I9fi9
Unhwsfty of l«nobof urt«na-OhampaJ
fln
College of Commerce and Business Administration
Bureau of Economic and Business Research
University of Illinois Urbana-Champaign
BEBR
FACULTY WORKING PAPER NO. 89-1577
College of Commerce and Business Administration
University of Illinois at Urbana-Champaign
June 1989
A Contingency Model of the Adoption ofthe Multidivisional Organization
Joseph T. Mahoney , Assistant ProfessorDepartment of Business Administration
A CONTINGENCY MODEL OF THE ADOPTION OF
THE MULTIDIVISIONAL ORGANIZATION
This paper examines the proposition that the multidivisional
structure is determined by both power and efficiency imperatives.
It is theorized that combining the coalitional power and
information-processing perspectives of organizational choice
improves predictive power. The theory is tested on 291 Fortune
500 firms. The results largely confirm theoretical expectations.
INTRODUCTION
A stream of research in strategic management has considered
the adoption of the multidivisional (M-form) as an adaptive
response to the problems of bounded rationality (an information-
processing imperative) and opportunism of organizational members
where control and auditing systems are inadeguate to mitigate the
agency problem of the separation of ownership and control
(Chandler, 1962; Williamson, 1975). Another stream of research
has considered organizational decisions from a coalitional power
perspective (Cyert and March, 1963; Pfeffer and Salancik, 1978).
While many researchers advocate the combining of efficiency and
power frameworks (Goldberg, 1980; Jemison, 1981; Lindblom, 1977;
Ulrich and Barney, 1984) there has been little empirical work
(with the notable exception of Palmer, et. al. (1987)) that
combines these theoretical perspectives.
From the efficiency perspective, the M-form has an
information-processing advantage relative to the large functional
enterprise. The functional form is subject to cumulative control
loss and a transformation of strategic formulation. Loss of
control results from serial reproduction loss as fragmentary or
erroneous information moves up and instructions are inadeguately
operationalized as they move down the hierarchy. In addition
lower level managers may intentionally falsify information to
their advantage (Williamson, 1970)
.
Strategy formulation may be altered as expansion of the
functional form ultimately overwhelms the ability of the top
level managers to provide corporate planning decisions and daily
coordination of operations (Galbraith, 1973; Mintzberg, 1979).
The M-form is viewed from the efficiency perspective as an
institutional response to problems of interdependence, subgoal
pursuit and confounding of strategic and operating decisions
(Williamson, 1975)
.
In addition to efficiency explanations for the emergence of
the multidivisional organization, an explanation based on the
power of coalitions in the organization has been proposed
(Pfeffer and Salancik, 1978) . Palmer, et. al. (1987) argue that
the conflict between top management and two types of external
ownership interests, family and financial institutions, may
influence the organizational choice.
Ultimately, the usefulness of these perspectives must be
determined empirically. However, following Palmer, et. al.
(1987), the model does not attend to the reductionist agenda of
choosing one perspective over another. Rather, the paper
maintains the general view that theoretical pluralism is a
legitimate methodology which increases empirical content (Bowman,
1989; Jemison, 1981), and specifically suggests that the
multidivisional organization may be best understood by
incorporating economic, administrative and power perspectives
(Bettis and Prahalad, 1983).
THE RESEARCH MODEL: A CONTINGENCY MODEL OF THE M-FORM
The multidivisional model tested is a contingency model for
predicting the likelihood of the adoption of the multidivisional
organization. The model analyzes the effects of firm size, firm
strategy (diversification) , environmental uncertainty and
coalitional power (Palmer, et . al. , 1987) on organizational
choice.
A well-grounded theoretical perspective is that structure
follows strategy (Chandler, 1962) and that in particular the
multidivisional structure follows the strategy of diversification
(Channon, 1973; Chenhall, 1979; Dyas & Thanheiser, 1976; Rumelt,
1974; Paven, 1976; Suzuki, 1980). The strategy of diversifica-
tion whether motivated by a resource-based imperative
(Penrose, 1959; Rubin, 1973; Wernerfelt, 1984); to obtain
technological capability (Nelson and Winter, 1982; Teece, 1982);
for financial reasons (Bowman, 1980; Jensen and Ruback, 1983;
Song, 1983), for managerial reasons (Amihud and Lev, 1981;
Mueller, 1969); to achieve synergies (Ansoff, 1965; Baumol,
Panzar and Willig, 1982) ; to reduce dependencies (Pfeffer and
Salancik, 1978) ; to reduce transaction costs (Williamson, 1985)
;
to utilize slack capacity (Chandler, 1977) ; or to increase market
power (Scherer, 1980) leads to problems of accountability,
control, and coordination.
A diversity of product lines tend to overload the decision
process of centralized organizations (Galbraith, 1977) . The
reorganization from the functional to the M-form attenuates
information overload problems. The M-form structure constitutes
a near-decomposable system to mitigate bounded rationality
constraints (Simon, 1962). The total system of decisions are
factored into "loosely coupled" subsystems (Weick, 1976)
.
An ideal multidivisional involves the following: (1)
Identification of separate economic activities and in particular
a separation of strategic and operating functions; (2)
Constructing quasi-autonomous divisions where profitability is
observable and measurable; (3) Monitoring the efficiency of each
division by a specialized corporate staff; (4) Awarding
incentives to promote profit-seeking behavior; (5) Allocating
cash flows to high yield uses; (6) Performing strategic planning
(Hill and Hoskisson, 1987; Williamson, 1975). Arguably, the
most important function of ^the M-form are the creation of its own
miniature capital market to achieve an efficient allocation of
capital (Heflebower, 1960; Williamson, 1970), and the attenua-
tion of bounded rationality and opportunistic behavior (William-
son, 1985)
.
The multidivisional form may also mitigate the agency
problem of the separation of ownership from control (Berle and
Means, 1932) since internal auditing and control systems
installed by the M-form overcome problems of asymmetric informa-
tion. Several studies support the M-form hypothesis that
multidivisionals, by attenuating the "information impactedness"
problem between corporate, business, and functional units,
increase profitability (Armour and Teece, 1978; Burton and Obel,
1980, 1988; Hill, 1985; Hoskisson and Galbraith, 1985; Teece,
1981; Thompson, 1981). However, a few studies do not support
the M-form hypothesis (Cable and Dirrheimer, 1983; Cable and
Yosuki, 1985; Harris, 1983), while others suggest a contingency
theory for the advantages of the M-form (Hill, 1988; Hoskisson,
1987) .
Hoskisson and Hitt (1987) suggest that even on theoretical
grounds, the M-form does not completely solve the agency problem
as the highly diversified multidivisional leads to a focus on
short-term profitability (Hayes and Abernathy, 1980; Loescher,
1984) . This latter group of studies questions whether the
multidivisional is an unequivocally superior organizational form.
If the contingency paradigm is correct (Galbraith, 1973;
Thompson, 19 67) then the M-form needs to be linked with the
interactive effects of efficiency and power variables in
predicting (and prescribing) organizational form.
Efficiency Perspective
From the Chandler-Williamson efficiency perspective, the
model tests whether diversification increases the likelihood of
the adoption of the M-form (HI) . Also, geographic dispersion
is expected to increase coordination and control problems, and
consequently is predicted to increase the likelihood of the
adoption of the M-form (H2) . Grinyer, Yasai-Ardekani and Al-
Bazzaz (1980) found this relationship positive and statistically
significant. A model which tests the separate effects of
diversification and geographic dispersion on organizational form
must also take into account the impact of an increase in
diversification increasing the geographic dispersion of the
enterprise (H3)
.
The generalizations of diversity - structure linkages must
be qualified by consideration of size. Self-contained product
divisions may be too small to have their own marketing, research,
or production department. Williamson (1975) argues that
increased size leads to the possibility of control loss within
the centralized organization and is an important variable in
determining organizational form. In contrast to Williamson's
theoretical perspective, Stopford and Wells (1972) argue that
absolute size by itself does not have a direct relationship with
(divisionalized) structure, that it is diversity that induces
divisionalization. Thus, the model needs to test the hypothesis
that increased size induces the adoption of the M-form structure
(H4) or whether an increase in size (capacity) leads to an
increase in diversity and/or geographic dispersion which results
in the M-form (Donaldson 1982, 1986) (H5) . Grinyer and Yasai-
Ardekani (1981) found that size exerts a direct causal influence
towards adopting the M-form. However, Donaldson (1982) found
that the association between size and the use of the M-form
disappeared when industrial diversity is controlled in partial
correlations. In both Donaldson's study and in our sample, only
Fortune 500 firms are considered so that the importance of size
may be under-estimated.
A major impediment to divisionalization is the existence of
a common technical system that cannot be segmented. Chandler
(1962) asserts a technological rationale for determining which
industries one may find diversification and ultimately the
multidivisional form. Industries that did not accept the M-form
structure were: [A] Copper and Nickel; [B] Steel; [C] Aluminum;
and [D] Materials {firms in these industries we shall designate
as METMAT} . Industries that only partially accepted the M-form:
[A] Petroleum companies; [B] Processors of agricultural products
{PETAGR}. Industries that widely accepted the M-form: [A]
Electrical and Electronics; [B] Power machinery and Automobiles;
[C] Chemicals (ELMACHEM) .
An aluminum producer despite large sales, a diversity of
customers and a variety of end products may be forced to retain a
functional structure because it can only afford one smelter.
Thus, it is not surprising to find that the aluminum, copper,
nickel and steel industries have been among those which have been
late to adopt the M-form (Chandler 1962) . Technologies with low
applications (steel, metal industries) imply that diversifica-
tion will be low and conseguently that the M-form will not be
adopted (H6) . Conversely, technologies with an abundance of
applications (electronics, chemicals, power machinery) imply
diversification and conseguently the likelihood of the adoption
of the M-form is expected to be much higher than petroleum and
agricultural firms (H7) .
Coalitional Power Perspective
In addition to the economic explanations for the emergence
of the multidivisional form, an explanation based on the power of
coalitions in the organization has been articulated by Cyert and
March (1963) and by Pfeffer and Salancik (1978). A conflict may
develop between top management and two types of external
ownership interests (families and financial institutions) which
influences organizational choice. In particular the model
considers whether family-dominated firms prefer centralized
control of operations and have a direct negative effect on the
likelihood that firms choose the M-form (H8) . Also, it is
hypothesized that family-dominated firms may resist diversifica-
tion (which dilutes their ownership and control over the firm)
and similarly may resist geographic dispersion (H9) . Several
case studies have observed that family-dominated firms tend to
resist adoption of the M-form (Chandler (1962) ; Channon (1973) ;
Pavan (1976)). Furthermore, Channon (1973) found that family-
controlled companies proved to be less diversified than the non-
family-controlled companies.
Palmer et al. (1987) hypothesized that bank-dominated firms
will be slow to adopt the M-form because the M-form threatens the
demand for the economy-wide investment information and expertise
of the banks. Since banks are in competition with the large
multidivisional ("a mini-bank"), the hypothesis is that bank
domination will imply a direct negative effect on the M-form
(H10) . Also, to the extent that bank dominated firms are less
diversified and less geographically dispersed, the indirect
effects will also lead to a lower likelihood of the adoption of
the M-form (Hll)
.
Finally, concerning life-cycles of the organization, we test
whether older firms due to structural inertia have a direct
negative effect on the adoption of the M-form (Fligstein, 1985;
8
Hannan and Freeman, 1984) (H12) . On the other hand, older firms
may be larger and may pursue geographic dispersion and diversity
which would lead to a positive effect (Chenhall, 1984) on the
choice of the M-form (H13) .
METHODS
A sample of 325 of the 500 largest U.S. industrials in 1965
were selected and classified along functional or multidivisional
lines. Missing data reduced the sample to 291 in all analyses.
The year 1965 was chosen because there were still a significant
number of functional organizations remaining in the Fortune 500.
Between 1966-1971, many of the remaining F-form structures became
M-form organizations (Bhargava, 1972, Hoskisson, 1987). This
surge in the diffusion process warrants closer scrutiny
(Mahajan, Sharna, and Bettis, 1987) . In the sample 194 firms
(2/3) were classified as multidivisional and 97 firms (1/3) were
classified as functional. Seven previous works were used to
validate the classifications (Armour and Teece, 1978; Bhargava,
1972; Chandler, 1962; Harris, 1983; Palmer, et. al. (1987);
Rumelt, 1974; Teece, 1981).
Of the 291 firms in the sample, 139 were classified by
Palmer et. al. (1987) and 12 of the 139 (8.6%) were inconsistent
with my classification. Of the 291 firms, 173 were classified
by Rumelt and 10 of the 173 (5.8%) were inconsistent with my
classification. Consistency of classification of organizational
form made independently by several researchers increases validity
and replicability.
Geographic dispersion is measured by three proxies: (1) The
number of geographically separate plants; (2) The number of
cities in which firm's plants operated; (3) The number of states
in which firm's plants operated. Data on the location of each
corporation's plants and on the industries in which they produced
were obtained from the Fortune 500 Plant and Product Directory,
1966.
Due to the large sample size, I chose to utilize SIC-based
measures of diversification, rather than Rumelt's classification
scheme. Montgomery (1982) found that the 2-digit, 3-digit, and
with one exception at the 4-digit level, SIC-based measures of
diversification (such as the Berry-Herf indahl index) increase
consistently with the strategy categories. Thus, there is a high
degree of correspondence between the continuous and categorical
measures. Montgomery noted that the Berry-Herf indahl measure is
particularly well suited for large sample cross-sectional
analysis. Diversification indices are simple, easy to compute,
objective, and replicable.
Several SIC measures have been articulated in the
literature. The proxy used by Palmer et. al. (1987) was a
product count measure. The analyses here will consider both
product count (Gort, 1962; Rhoades, 1973) and Berry (1975)
measures of diversification.
A drawback of the product count measure is that undue weight
is given to minor activities and the SIC classifications are
somewhat arbitrary. Merely counting product lines exaggerates
the overall significance of diversification since most firm's
10
product volume distributions are highly skewed, with a few
product lines accounting for the bulk of sales or employment
while numerous other lines are relatively small. A firm, 99% of
whose sales were accounted for by a single 5-digit product is
hardly diversified regardless of the number of 4-digit industries
represented in by the remaining one percent.
The Berry index corresponds to the Hirschman-Herf indahl
B = 1 - £index: ^* 2
l* I1
thwhere P = ratio of the firm's output in the i industry to the
i
total output. This measure of diversification considers not
only the number of industries in which a firm is active, but also
the distribution of the firm's production activity among those
industries. A firm with 99 percent of its output accounted for
by a single 4-digit product is not diversified regardless of the
number of 3-digit industries represented by the remaining one
percent. On the other hand, a firm with its productive activity
egually divided among four 3-digit industries is likely to be
"diversified", even if no more than four 4-digit products are
involved.
The index takes on a value of for a specialized firm
acting in a single industry and approaches unity when a firm
produces egually in a large number of industries = (1 - 1/N)
,
where N= number of industries in which it is active. The index
is comparatively insensitive to minor secondary activities. I
maintain that this measure is in fact desirable for studying the
effects of diversification on organizational change. The
11
empirical analysis considers the Berry-Herf indahl index across 2-
digits, 3-digits and 4-digits. The average Berry index for 1965
across 2-digits for my 291 firm sample was .406, the average
across 4-digits was .679.
The size of the Fortune 500 firm is measured by four alter-
native proxies: (1) Sales; (2) Assets; (3) Invested Capital;
and (4) Employees. Corporate Age is measured by the number of
years (in decades) between 1965 and the year the firm was
incorporated. The year of incorporation was obtained from
Moody's Handbook of Common Stocks. The primary industry in
which each firm produced was obtained from the U.S. Bureau of the
Census (1977) .
Burch's (1972) study of the largest 500 U.S. industrial
corporations in 1966 was used to measure dominance by family
coalition. Three categories suggested by McEachern (1975) were
used:
"Free of family influence" — if no identifiable group of relatedpeople owned more than 4% of their stock;
"Family owned" (FOWN =1) — if more than 4% of their stock wasowned by group of related people, none of whom were inside boardmembers, otherwise FOWN=0;
"Family owned and controlled" (FOAC =1) — if more than 4% oftheir stock was owned by a group of related people, at least oneof whom was an inside board member, otherwise FOAC=0.
The U.S. Congress House Committee on Banking and Currency,
Pattman Subcommittee on Domestic Finance (1968) was used to
measure dominance by bank coalition. This volume lists the
amount of stock and number of board seats 49 large financial
12
institutions held in the largest 500 U.S. industrial corporations
in 1966:
If no bank or combination of banks owned at least 5% of a firm'sstock, it was considered "free of bank influence";
If more than 5% of a firm's stock was owned by a bank or group ofbanks, but none of the board seats were held by representativesof these institutions, the firm was considered "bank owned"(BOWN = 1; O otherwise)
;
If more than 5% of a firm's outstanding common and preferredstock (with partial or full voting rights) was owned by a bank orgroup of banks and one or more of its board seats was held by arepresentative of this bank or group of banks the firm wasconsidered "bank owned and controlled" (BOAC =1; otherwise);
A summary of the variables used in the study is given in Table 1.
Table 2 gives details of the means, standard deviations, and
correlations for all the variables. There is no apparent
problem of multicollinearity , and the correlations give strong
indications that the hypotheses generated earlier are on target.
Insert Tables 1 and 2 about here
A TEST OF THE MULTIDIVISIONAL THEORY
Hypotheses were tested by estimating a system of structural
eguations:
(1) MF = f [LNST, BDIV4, LNEMPL, METMAT, PETAGR, AGE, FOWNFOAC, BOWN, BOAC
]
(2) LNST = f [BDIV4, LNEMPL, METMAT, PETAGR, AGE, FOWNFOAC, BOWN, BOAC
]
(3) BDIV4 = f [LNEMPL, METMAT, PETAGR, AGE, FOWN, FOAC,BOWN, BOAC ]
13
The second and third equations are estimated using ordinary-least
squares linear regression (Tables 4 and 5) . Because of the
binary dependent variable (MF =0, or MF =1), the logistic
response function is used to represent the impact of the effects
on the probability of becoming multidivisional in the first
equation (Table 3) . The logit model allows the use of
categorical or discrete variables for both dependent and
independent variables. Since the model contains qualitative
independent variables, logistic regression is chosen over
discriminant analysis (Press and Wilson, 1978) .
Letting X,X ,X ,... X =X1J 2J 3J 10J J
stand for the 10 factors described above for subject J, we have:
io \o
P(MF = 1 | X ) = exp ( B + £ B X ) / 1 + exp ( B + ^ B X )
J J i=l i iJ i=l i iJ
where MF = { if the enterprise is not multidivisionalJ 1 if the enterprise is multidivisional
P (MF = 1| X ) is the probability that a firm with company andJ J
market characteristics X uses a multidivisional structure.J
Taking the natural logarithm of both sides of the equation yields
the linear relation between the factors and the logit or log odds
ratio:
J J J J i=l i iJLN[P(MF = 1 | X )/l-P (MF = l\ X ) =B + £» B X
14
The coefficients were estimated by maximizing the likelihood
function:
«TT» lMFT 1-MF,
L (MFI
X ; B ) = TT P (MF = 1 I X ) * (1- P (MF = 1 X )^
J J J=l J ' J J J
where N = the 291 firms on which the data have been collected.
A noteworthy feature of this model is that even though the
dependent variable is binary, the model's predictions are not.
Rather, the model's predictions are estimates of the probability
of taking on the value of 1 (rather than 0) . Maximization of
the likelihood function was accomplished with the Gauss-Newton
nonlinear least squares method.
Insert Tables 3, 4, and 5 about here
To test the hypothesis that the explanatory variables have no
impact on the choice of probabilities P , that is, thei
B = B = . . . B =12 10
the test statistic is -2 [ In 1 ( @ ) - In 1 (w)]
where 1 ( § ) is the value of the likelihood function evaluated
A.
at the maximum likelihood estimates and 1 (w) is the maximum
value of the likelihood function under the hypothesis that
B =B =B =0. If the hypothesis is true, then asymp-1 2 10
totically, the test statistic has a chi-square distribution with
15
(K-l) degrees of freedom. From our logit regression, the value
of the test statistic is -2 (126.91 - 185.23) = 116.624. The
chi-sguare with 10 degrees of freedom at the one percent level of
significance eguals 25.188, so that we can reject the hypothesis
that B=B = ... B =0. A related summary measure12 10
2A
is the McFadden R computed as 1 - In 1 ( § ) / In 1 (w)
= 1 - 126.91 / 185.23 = .31482. This measure has value zero
when B = B = . . . B =0 and value 1 when the model is a12 10
perfect predictor. This measure is analogous to the coefficient
2
of determination R in linear regression models.
RESULTS
As Table 6 below shows, the results support the economic
explanation of the M-form.
Insert Table 6 about here
An increase in diversification, as measured by the Berry-
Herfindahl 4-digit index, significantly increases the likelihood
that the enterprises use the M-form in support of HI. The
results were robust across the seven diversification measures
used. The results were also robust using probit analysis.
Geographic dispersion, as measured by the log of the number
of states in which the enterprise had plants, significantly
16
(p <.01) increases the likelihood that the enterprises use the M-
form, supporting H2 . The results also hold when the log of
cities or the log of plants were used as proxies for geographic
dispersion. From the OLS regressions, diversification signif-
icantly increased geographic dispersion in support of H3
.
While Palmer et al. (1987) found a slightly negative
relationship between size (measured by the log of employees) and
the likelihood of the enterprises adopting the multidivisional
structure, in our study, an increase in size (LNEMPL) was
positively associated with the M-form, but did not increase the
likelihood of the M-form at a statistically significant level.
This supports the Stopford and Wells (1972) argument that
diversification, rather than size per se has an influence on
organizational form. This conclusion from the data however is a
tentative one. The result is not robust across size measures
(Kimberly, 1976) . When size is measured by the log of assets
for example, while all other regression results hold, the size
variable is positive and significant (p <.10), supporting H4
.
That an increase in size leads to an increase in the
likelihood of the M-form is suggested by Williamson (1975) and is
consistent with the empirical results of Grinyer and Yasai-
Ardekani (1981) . Since the sample in the study is restricted to
the Fortune 500, this significant result should apply a fortiori
to a general enterprise population sample. Further empirical
work is required to determine the influence of size on
organizational form. The OLS equations indicate that increased
size also induces increased diversification and increased
17
geographic dispersion in support of H5 . This result was robust
across size measures, dispersion measures and diversification
measures.
Consistent with Chandler (1962), the industries associated
with high capital requirements and low technically driven diver-
sification were significantly less likely to adopt the M-form.
The logit analysis indicates that the metals and materials firms
were significantly (p <.05) less likely to adopt the M-form which
supports Chandler (1962, Ch.7). The Palmer et al. (1987) study
on the other hand, did not support Chandler's findings that the
metals and materials firms were less likely to adopt the M-form.
Besides using a Berry diversification measure and a larger sample
size, the discrepancy between their results and ours is partly
due to the discrepancy in the classification of M-form and F-
form. For example, they classified such firms as Kennecott
Copper Corp. and Republic Steel as multidivisional while several1
other independent researchers have classified them as functional.
The petroleum and agricultural firms (p <.01) were less
likely to utilize the M-form than the enterprises whose primary
industry was chemical, machinery, or electrical in support of H7.
The petroleum and agricultural firms were also significantly less
diversified but they were significantly more geographically
dispersed than the chemical, electrical, and machinery firms.
1 This criticism of the Palmer et. al. (1987) study is notintended to be contentious. In fact, the author regards thepaper as an exemplar of scholarly work, based on well-groundedtheoretical and empirical work which is completely replicable.This paper in fact bears out the robustness of the model in termsof a different sample and alternative proxies for variables. Thepaper highlights those areas where caution in predictions need tobe exercised.
18
The results also support the political coalition view of the
firm (Pfeffer and Salancik, 1978) . Those firms that were family
dominated (FOWN, FOAC) were significantly (P <.10; p <.01) less
likely to adopt the multidivisional structure in support of H8.
The family-dominated firms FOWN also diversified significantly
less (p <.05) in support of H9 . Family coalitions resist diver-
sification because it threatens their ownership and control. If
diversification via acguisitions is financed by debt, the power
of banks in firm's long-run decisions increase. If acguisitions
are financed by issuing new stock, then the holding of family
members are diluted and outside managers are reguired which
reduces the power of family members. In contrast to the Palmer
et al. (1987) study, our results indicate no effect of family-
dominance (FOWN, FOAC) on geographic dispersion.
Bank-dominated firms (BOWN, BOAC) were less likely to adopt
the M-form in support of H10. However, only the BOWN enterprises
were significantly (p <.10) less likely to adopt the M-form.
Palmer et al. (1987: 39) suggest a possible rationale for this
result:
Banks may not discourage firms from adopting the M-form asvigorously when they own and control (as opposed to onlyown) them, because they are in a position to insure that theadoption of this form does not allow a firm to internalizethe capital market. By placing representatives on the board(and perhaps the finance committee) , banks may be able tocontrol a firm's capital allocation process, when banks areonly the dominant stockholders in a firm, they may not beable to exercise such influence on a regular basis.
Also, bank-owned and controlled firms were significantly (p <.05)
less dispersed geographically. However, there was no effect of
bank domination on diversification in contradiction to Hll.
19
Little support was found for the organizational variant of
the ecological approach. Although AGE was negatively associated
with adoption of the M-form suggesting a structural inertia
effect, the effect was not statistically significant in contra-
diction to H12 . The age of the enterprise also had no effect on
geographic dispersion. However, in contrast to Palmer, et. al
(1987), the results indicated that a firm's increase in age leads
to a significant (p <.05) increase in diversification in support
Of H13.
DISCUSSION
This eclectic model considers traditional industrial
economics variables (such as the Berry-Herfindahl index and the
influence of the primary industry) , and a coalitional view of the
firm (Cyert & March, 1963; Pfeffer and Salancik, 1978), where
family-dominance and bank dominance are important factors in
explaining and predicting an enterprise's strategy and structure.
The model significantly improves the explanation and prediction
of organizational form.
Although some of the results were not consistent with
Palmer, et. al. (1987), the overall conclusion is that the model
proved guite robust to changes in sample and proxies, for this
time period. A guestion to be addressed in future research is:
How well does the model predict organizational form for later
(or earlier) time periods? The model presented stands up guite
well to the criteria of multiple connectedness and replicability
.
A well-grounded theoretical and empirical literature suggests
20
that the model is generalizable . Of course, this latter
assertion must be backed with the hard currency of further
empirical efforts.
Finally, it is submitted that the multidivisional paradigm
illustrates the central premises of the paper: (1) a synthesis of
efficiency and power perspectives is a viable research program;
and (2) theoretical pluralism increases empirical content and
should be valued by those concerned with progress in the emerging
field of strategic management (Bowman, 1989, Huff, 1981).
21
TABLE 1
Summary of variables used in regressions presented
Variable
MF
LNST
BDIV4
LNEMPL
METMAT
Description
PETAGR
= 1 if firm is multidivisional= if functional
Natural log of the number ofstates that the enterprise hadplants (*)
Berry-Herf indahl 4-digit meas-ure of diversification (**)
Natural log of the number ofemployees of the enterprise( ***)
= 1 if enterprise's primaryindustry is in metals ormaterials
= otherwise
= 1 if enterprise's primaryindustry is in petroleum oragriculture
= otherwise
FOWN
FOAC
Family-owned
Family-owned and controlled(Defined in Methods section)
BOWN
BOAC
AGE
Bank-owned
Bank-owned and controlled(Defined in Methods section)
(1965- Year of Incorporation) /10
* = two other measures of geographic dispersion used
** = six other diversification measures used
*** = three other size measures used
22
TABLE 2
Means, Standard Deviations and Correlations
Mean SD 1234567 89 10
1 MF .67 .47
2 LNST 2.18 .81 .39
3 BDIV4 .68 .24 .42 .41
4 LNEMPL 9.83 .97 .32 .39 .31
5 METMAT .15 .35 -.12 .01 -.06 -.04
6 PETAGR .22 .41 -.25 .10 -.32 -.16 -.22
7 AGE 5.78 2.26 .06 .19 .18 .21 .04 .05
8 FOWN .09 .29 -.09 .05 -.11 .03 .08 .01 -.02
9 FOAC .43 .50 -.25 -.15 -.09 -.23 -.01 .06 -.02 -.28
10 BOWN .09 .28 -.11 -.03 .01 .02 -.03 .03 -.06 .05 .10
11 BOAC .12 .35 -.06 -.09 .01 .09 .08 -.05 .01 -.07 -.01 -.13
Pearson product-moment correlations are used when both variables
are continuous. Spearman rank-order correlations are used when
at least one variable is categorical.
23
TABLE 3
Logit Regression Dependent Variable: MF
VARIABLE NAME ESTIMATED COEFFICIENT T-RATIO
LNST 1.0158 4.1099 ***
BDIV4 2.2345 2.9832 **
LNEMPL 0.3495 1.6469
METMAT -1.2479 -2.7125 **
PETAGR -1.5206 -3.6192 ***
AGE -0.0812 -1.1025
FOWN -1.0136 -1.8899 *
FOAC -1.0809 -3.1613 **
BOWN -0.8892 -2.0608 *
BOAC -0.6317 -1.1485
Constant -4.4476 -2.2536
* =(p <.10) ** = (p <.05) *** = (p <.01)
Log Likelihood (0) = -185.23
Log Likelihood Function = -126.91
2
McFadden R = .31482
2
Craig-Uhler R = .45859
The number of correct predictions from the model was 233. The
percentage of correct predictions then was 80.07 percent.
24
TABLE 4
OLS Regression
Variable Name
BDIV4
LNEMPL
METMAT
PETAGR
AGE
FOWN
FOAC
BOWN
BOAC
Constant
F value
Dependent Variable LNST
Estimated Coefficient
1.329
0.262
0.180
0.543
0.015
0.240
-0.093
-0.143
-0.365
-1.454
16.2
T Ratio
7.28 ***
5.74 ***
1.55
5.22 ***
0.84
1.64
-1.08
-1.24
-2.59 **
R = .342 * = (p <.10) ** = (p < .05) *** = (p < .01)
25
TABLE 5
OLS Regression
Variable Name
LNEMPL
METMAT
PETAGR
AGE
FOWN
FOAC
BOWN
BOAC
Constant
F value 8.67
Dependent Variable BDIV4
Estimated Coefficient
.056
-.067
-.157
.015
-.106
-.022
.026
-.008
.103
T Ratio
3.87 ***
-1.77
-4.80 ***
2.50 ***
-2.23 **
-0.80
0.69
-0.18
R = .198 * == (P <-10) ** - (P <-05) *** == (P < -01)
26
TABLE 6
MULTIDIVISIONAL MODEL
HYPOTHESIS RESULT
HI: Diversification induces theadoption of the M-form
ACCEPT(p<.05)
H2: Geographic dispersion inducesthe adoption of the M-form
ACCEPT(p<.01)
H3 : Diversification increasesgeographic dispersion
ACCEPT(p<.01)
H4 : Increased size induces theadoption of the M-form
REJECT
H5: (a) Increased size leads toan increase in diversification
(b) Increased size results inan increase in geographicdispersion
ACCEPT(p<.01)
ACCEPT(p<.01)
H6: Technologies with low applica-tions (Metals & Materials) willhave low diversification & thusadoption of the M-form is lesslikely
ACCEPT(p<.05)
H7 : Technologies with high applica-tions such as electronics, chem-icals & power machinery will adoptthe M-form with a higher probabil-ity than petroleum and agricultur-al firms
ACCEPT(p<.01)
H8 : Family-dominated firms prefer cen-tralized control and will have adirect negative effect on the like-lihood that firms choose the M-form
ACCEPTFOWN (p<.10)FOAC (p<.05)
27
H9 : (a) Family-owned firms may resistdiversification, which dilutes theirownership and control over the firm
ACCEPTFOWN (p<.10) only
(b) Family-dominated firms resistgeographic dispersion
REJECTFOWN and FOAC
H10: Bank domination will have a directnegative effect on adoption of theM-form
ACCEPTBOWN (p<.10) only
Hll: (a) Bank- dominated firms are ex-pected to be less diversified
REJECTBOWN and BOAC
(b) Bank-dominated firms are ex-pected to be less geographicallydispersed
ACCEPTBOAC (P<.10) only
H12: Older firms, due to structural in-ertia, are expected to have a neg-ative effect on the adoption of theM-form
REJECT
H13: (a) Older firms may pursue greatergeographic dispersion
REJECT
(b) Older firms are expected to havegreater diversification
ACCEPT(p<.05)
28
ACKNOWLEDGMENT
The author wishes to thank Edward H. Bowman, Irene Duhaime, Jerry
Goodstein, Bob Hill, Anne Huff, Paula Rechner, Ming-Je Tang, and
Howard Thomas for helpful comments on an earlier draft of the
paper and Bill Bogner for research assistance. Support from the
Reginald Jones Center is gratefully acknowledged.
REFERENCES
Allen, Stephen A. , "Understanding Reorganizations of Division-alized Companies," Academy of Management Journal , Vol.22 (1979),pp. 641-671.
Amihud, Yakov, and Baruch Lev, "Risk Reduction as a ManagerialMotive for Conglomerate Mergers," Bell Journal of Economics ,
(1981) , pp. 605-617.
Ansoff, H. Igor, Corporate Strategy . New York: McGraw Hill,1965.
Ansoff, H. I., and R. G. Brandenburg, "A Language for Organiza-tional Design: Part II," Management Science . Vol.17 (1971),pp. B 717-731.
Armour, Henry Ogden and David J. Teece, "Organizational Struc-ture and Economic Performance: A Test of the MultidivisionalHypothesis," Bell Journal of Economics . Vol.9 (1978), pp. 106-122.
Baumol, William J., John C. Panzar and Robert D. Willig,Contestable Markets and The Theory of Industry Structure . NewYork: Harcourt Brace Jovanovich, 1982.
Berle, Adolf A., and Gardiner C. Means, The Modern Corporationand Private Property . New York: Macmillan, 1932.
Berg, Norman A. , "Whats Different About Conglomerate Manage-ment?," Harvard Business Review . Vol.47 (1969), pp. 112-120.
Berry, Charles H., Corporate Growth and Diversification .
Princeton, N. J. : Princeton University Press, 1975.
Bettis, Richard A. and C.K. Prahalad, "The Visible and the In-visible Hand: Resource Allocation in the Industrial Sector,"Strategic Management Journal . Vol.4 (1983), pp. 27-43.
Bhargava, Narottam, The Impact of Organization Form on the Firm:Experience of 1920-1970 . University of Pennsylvania, DoctoralDissertation, 1972.
Bowman, Edward H. , "A Risk/Return Paradox for Strategic Manage-ment," Sloan Management Review . Vol.21 (Spring 1980), pp. 17-31.
Bowman, Edward H. , "Strategy Changes," Reginald H. Jones Centerfor Management Policy, Strategy and Organization, Wharton School.Philadelphia, Pa.: University of Pennsylvania (1988-02).
Bowman, Edward H. , "Strategy Changes: Possible Worlds and ActualMinds," Perspectives on Strategic Management , James Fredrickson,editor, Ballinger Company, forthcoming, 1989.
Burch, Phillip, The Managerial Revolution Reassessed . Lexington,Ma.: D.C. Heath, 1972.
29
Burton, Richard M. and Borge Obel, "A Computer Simulation Testfor the M-form Hypothesis," Administrative Science Quarterly .
Vol.25 (1980), pp. 457-466.
Burton, Richard M. and Borge Obel, "Opportunism, Incentives, andthe M-form Hypothesis," Journal of Economic Behavior andOrganization . Vol.10 (1988), pp. 99-119.
Cable, J. R. and M. J. Dirrheimer, "Markets and Hierarchies: AnEmpirical Test of the Multidivisional Hypothesis in WestGermany," International Journal of Industrial Organization .
Vol.1 (1983), pp. 43-62.
Cable, J. R. and H. Yosuki, "International Organization,Business Groups and Corporate Performance: An Empirical Test ofthe Multidivisional Hypothesis in Japan," International Journalof Industrial Organization . Vol.3 (1985), pp. 401-420.
Caves, Richard E. , "Industrial Organization, Corporate Strategyand Structure," Journal of Economic Literature . Vol. 18 (March1980) , pp. 64-92.
Chandler, Alfred D. , Jr., Strategy and Structure: Chapters in theHistory of The Industrial Enterprise . Cambridge, Mass. : MITPress, 1962.
Chandler, Alfred D. , Jr. , The Visible Hand: The ManagerialRevolution in American Business . Cambridge, Mass.: BelknapPress, 1977.
Chandler, Alfred D. Jr. and Herman Daems, (ed.), ManagerialHierarchies . Cambridge: Harvard University Press, 1980.
Channon, Derek F. , Strategy and Structure in British Enterprise .
Boston, MA.: Harvard University Press, 1973.
Chenhall, Robert H. , "Diversification Within Australian Manu-facturing Enterprises," Journal of management Studies . VOL 21(1984) , pp. 23-60.
Cyert, Richard M. , and James G. March, A Behavioral Theory ofthe Firm . Englewood Cliffs, N.J.: Prentice-Hall, Inc., 1963.
Didrichsen, Jon, "The Development of Diversified andConglomerate Firms in the United States, 1920-1970," BusinessHistory Review . Vol.64 (Summer 1972), pp. 202-219.
Donaldson, Lex, "Divisionalization and Size: A Theoretical andEmpirical Critique," Organization Studies . Vol.3 (4) (1982), pp.321-337.
Donaldson, Lex, "The Interaction of Size and Diversification asa Determinant of Divisionalization: Grinyer Revisited," Organi-zation Studies . Vol.7 (4) (1986), PP. 367-379.
30
Dundas, Kenneth N. M. , and Peter R. Richardson, "Implementingthe Unrelated Product Strategy," Strategic Management Journal .
Vol.3 (1982), pp. 287-301.
Dyas, Gareth P., and Heinz T. Thanheiser, The Emerging EuropeanEnterprise . London: Macmillan & CO., 1976.
Edstrom, Anders, and Jay R. Galbraith, "Transfer of Managers asa Coordination and Control Strategy in Multinational Organiza-tions," Administrative Science Quarterly . Vol.22 (June 1977),pp. 248-263.
Egelhoff, William G. , "Strategy and Structure in MultinationalCorporations: An Information Processing Approach," Administra-tive Science Quarterly . Vol.27 (1982), pp. 435-458.
Fligstein, Neil, "The Spread of the Multidivisional Form: 1919-1979," American Sociological Review . Vol.50 (1985), pp. 377-391.
Fortune Plant and Product Directory, 19 66. New York: Fortune,Department of Marketing Research, 1966.
Fouraker, Lawrence, and John M. Stopford, "OrganizationalStructure and Multinational Strategy," Administrative ScienceQuarterly . Vol.13 (1968), pp. 47-64.
Frederickson, James W. , "The Strategic Decision Process andOrganizational Structure," Academy of Management Review . Vol.11(1986) , pp. 280-297.
Franko, Lawrence G., "The Move Toward a MultidivisionalStructure in European Organizations," Administrative ScienceQuarterly . (1974), pp. 493-506.
Galbraith, Jay, Designing Complex Organizations . Reading,Mass.: Addison-Wesley, 1973.
Galbraith, Jay, Organization Design . Reading, MA. : Addison-Wesley, 1977.
Galbraith, Jay R. , and Daniel A. Nathanson, Strategy Implementa-tion: The Role of Structure and Process . St. Paul, Minn.: WestPublishing, 1978.
Goldberg, Victor P., "Bridges Over Contested Terrain: Exploringthe Radical Account of the Employment Relationship," Journal ofEconomic Behavior and Organization . Vol.1 (1980), pp. 249-274.
Gort, Michael, Diversification and Integration in AmericanIndustry . Princeton, N. J. : Princeton University, 1962.
Govindarajan, Vijay, "Decentralization, Strategy, and Effect-iveness of Strategic Business Units in Multibusiness Organiza-tions," Academy of Management Review , Vol.11 (1986), pp. 844-856.
31
Grinyer, Peter H. , Masoud Yasai-Ardekani, and Shawki Al-Bazzaz,"Strategy and Structure, The Environment and FinancialPerformance in 48 United Kingdom Companies," Academy ofManagement Journal , Vol.23 (1980), pp. 193-220.
Grinyer, Peter H., and Massoud Yasai-Ardekani, "Strategy,Structure, Size, and Bureaucracy," Academy of ManagementJournal , Vol.24 (1981), pp. 471-486.
Grinyer, Peter H. , "Discussion Note: Divisionalization and Size:A Rejoinder," Organization Studies , Vol.3 (1982), pp. 339-350.
Hannan, Michael, and John Freeman, "Structural Inertia andOrganizational Change," American Sociological Review , Vol.49(1984) , pp. 149-164.
Harris, Barry C. , Organization: The Effect on Large Corpora-tions . Ann Arbor: UMI Research Press, 1983.
Hayes, Robert H. , and William J. Abernathy, "Managing Our Way toEconomic Decline," Harvard Business Review . Vol.58 (1980), pp.67-77.
Heflebower, Richard B. , "Observations on Decentralization inLarge Enterprises," Journal of Industrial Economics , Vol.9(1960) , pp. 7-22.
Herbert, Theodore T. , "Strategy and Multinational OrganizationStructure: An Interorganizational Relationship Perspective,"Academy of Management Review . Vol.9 (April 1984), pp. 259-270.
Hill, Charles W. L. , "Internal Organization and EnterprisePerformance: Some U.K. Evidence," Managerial and DecisionEconomics . Vol.6 (1985), pp. 210-216.
Hill, Charles W. L. , and J.F. Pickering, "Divisionalization,Decentralization, and Performance of Large U.K. Firms," Journalof Management Studies . Vol.23 (January 1986), pp. 26-50.
Hill, Charles W. L. , and Robert E. Hoskisson, "Strategy andStructure in the Multiproduct Firm," Academy of ManagementReview . Vol.12 (1987), pp. 331-341.
Hill, Charles W. L. "Internal Capital Market Controls andFinancial Performance in Multidivisional Firms," Journal ofIndustrial Economics . V01.37 (September 1988), pp. 67-83.
Hoskisson, Robert E., "Multidivisional Structure andPerformance: The Diversification Strategy Contingency," Academyof Management Journal . Vol.30 (December 1987), pp. 625-644.
Hoskisson, Robert E. , and Michael A. Hitt, "Strategic ControlSystems and Relative R&D Investment in Large Multiproduct Firms,"Strategic Management Journal . Vol.9 (1988), pp. 605-621.
32
Huff, Anne S., "Multilectic Methods of Inquiry," Human SystemsManagement . Vol.2 (1981), pp. 83-94.
Jemison, David B. , "The Importance of an Integrative Approach toStrategic Management Research," Academy of Management Review .
(October 1981), pp. 601-608.
Jennegren, L. Peter, "Decentralization in Organizations," inPaul Nystrom and William Starbuck, eds., Handbook of Organiza-tional Design . 2 vols., New York: Oxford University Press, 1981.
Jensen, Michael C. , and R. Ruback, "The Market for CorporateControl: The Scientific Evidence," Journal of FinancialEconomics . Vol. 11 (1983), pp. 5-50.
Johnson, H. Thomas, "Management Accounting in an Early Multi-divisional Organization: General Motors in the 1920's," BusinessHistory Review . Vol. 52 (Winter 1978), pp. 490-517.
Khandwalla, Pradip, N. , "Viable and Effective OrganizationalDesign of Firms," Academy of Management Journal . Vol. 16(3)(September 1973) , pp. 481-495.
Kimberly, John R. , "Organizational Size and the StructuralistPerspective: A Review, Critique, and Proposal," AdministrativeScience Quarterly . Vol.21 (December 1976), pp. 571-597.
Lindblom, Charles E. Politics and Markets . New York: BasicBooks, Inc. , 1977.
Loescher, Samuel M. , "Bureaucratic Measurement, Shuttling StockShares, and Shortened Time Horizons: Implications for EconomicGrowth," Quarterly Review of Economics and Business . Vol.2 4
(Winter 1984), pp. 8-23.
Mahajan, Vijay, Subbash Sharna, and Richard A. Bettis, "TheAdoption of the M-form Organizational Structure: A Test of theImitation Hypothesis," Southern Methodist University, November,1987.
March, James G. and Herbert A. Simon, Organizations . New York:John Wiley & Sons, Inc., 1958.
Margenson, Paul, "The Multidivisional Firm and Control over theWork Process," International Journal of Industrial Organization .
Vol.3 (1985), pp. 37-56.
McCloskey, Donald N. , "The Rhetoric of Economics," Journal ofEconomic Literature . Vol.21 (June, 1983), pp. 481-517.
McEachern, William, Managerial Control and Performance . Lexing-ton, MA.: D. C. Heath, 1975.
33
McFadden, Daniel, "Conditional Logit Analysis of QualitativeChoice Behavior," in P. Zarembka, ed. , Economic Theory andMathematical Economics . New York: Academic Press, 1974.
Miller, Danny, "Strategy Making and Structure: Analysis andImplications for Performance," Academy of Management Journal .
Vol.30 (March 1987), pp. 7-32.
Mintzberg, Henry, The Structuring of Organizations . EnglewoodCliffs: Prentice-Hall, 1979.
Montgomery, Cynthia A. , "The Measurement of Firm Diversifica-tion: Some New Empirical Evidence," Academy of ManagementJournal . Vol.25 (1982), pp. 299-307.
Mueller, Dennis C, "A Theory of Conglomerate Mergers,"Quarterly Journal of Economics (November 1969), pp. 644-659.
Nelson, Richard R. , and Sidney G. Winter, An Evolutionary Theoryof Economic Behavior and Capabilities . Cambridge, Mass.: HarvardUniversity Press, 1981.
Negandhi, Anant R. , and Bernard C. Reimann, "Correlates ofDecentralization: Closed and Open Systems Perspective," Academyof Management Journal . Vol. 16(4) (December 1973), pp. 570-582.
Palmer, Donald, Roger Friedland, P.Devereaux Jennings, andMelanie E. Powers, "The Economics and Politics of Structure: TheMultidivisional Form and the Large U.S. Corporation,"Administrative Science Quarterly . Vol.32 (1987), pp. 25-48.
Pavan, Robert J., "Strategy and Structure: The ItalianExperience," Journal of Economics and Business . Vol.28 (1976),pp. 254-260.
Penrose, Edith T., The Theory of the Growth of the Firm .
Oxford: Blackwell, 1959.
Pfeffer, Jeffrey, and Gerald Salancik, The External Control ofOrganizations . New York: Harper and Row, 1978.
Pfeffer, Jeffrey, Organizations and Organization Theory . Bos-ton: Pitman, 1982.
Pitts, Robert A., "Diversification Strategies and OrganizationalPolicies for Large Diversified Firms," Journal of Economics andBusiness . Vol.28 (1976), pp. 181-188.
Pitts, Robert A., "Strategies and Structures for Diversifica-tion," Academy of Management Journal . Vol.20 (1977), pp. 197-208.
Press, S. James, and Sandra Wilson, "Choosing Between LogisticRegression and Discriminant Analysis," Journal of the AmericanStatistical Association . V01.73 (December 1978), pp. 699-705.
34
Rhoades, Stephen A., "The Effects of Diversification on IndustryProfit Performance in 241 Manufacturing Industries, 1963,"Review of Economics and Statistics , Vol.55 (1973), pp. 146-155.
Rubin, Paul H. , "The Expansion of Firms," Journal of PoliticalEconomy . Vol.81 (July-August 1973), pp. 936-949.
Rumelt, Richard P. , Strategy, Structure, and Economic Perform-ance . Division of Research, Graduate School of Business Adminis-tration, Harvard University, 1974.
Scherer, F. M. , Industrial Market Structure and EconomicPerformance . Boston: Houghton Mifflin Company, 1980.
Simon, Herbert A., "The Architecture of Complexity," Proceed-ings of the American Philosophical Society . Vol.106 (December1962) , pp. 467-482.
Simon, Herbert A. , Administrative Behavior . New York: The Mac-millan Company, 1947 (3rd edition 1976)
.
Song, Jae H., "Diversifying Acquisitions and FinancialRelationships: Testing 1974-1976 Behavior," Strategic ManagementJournal . Vol.4 (1983), pp. 97-108.
Steer, Peter and John Cable, "Internal Organization and Profit:An Empirical Analysis of Large U.K. Companies," Journal ofIndustrial Economics . Vol.27 (September 1978), pp. 13-30.
Stopford, John M. and Louis T. Wells, Managing the MultinationalEnterprise . New York: Basic Books, Inc., 1972.
Suzuki, Y., "The Strategy and Structure of Top 100 JapaneseIndustrial Enterprises, 1950-1970," Strategic Management Journal .
Vol.1 (3) (July-September 1980), pp. 265-291.
Teece, David J., "The Diffusion of an Administrative Innova-tion," Management Science . Vol.26 (May 1980), pp. 464-470.
Teece, David J. , "Internal Organization and Economic Perform-ance: An Empirical Analysis of the Profitability of PrincipalFirms," Journal of Industrial Economics , Vol.30 (December 1981),pp. 173-199.
Teece, David J., "Towards an Economic Theory of the MultiproductFirm," Journal of Economic Behavior and Organization . Vol.3(1982) , pp. 39-63.
Thompson, James D. , Organizations in Action . New York: McGraw-Hill, 1967.
Ulrich, David, and Jay B. Barney, "Perspectives in Organiza-tions: Resource Dependency, Efficiency, and Population," Academyof Management Review , Vol.9 (1984), pp. 471-481.
35
U.S. Bureau of the Census "General Report on IndustrialOrganization," 1972 Enterprise Statistics . Part I, Washington,D. C. , Government Printing Office, 1977.
U.S. Congress House Committee on Banking and Currency, PattmanSubcommittee on Domestic Finance. 1968 Commercial Banks andTheir Trust Activities: Emerging Influence on the AmericanEconomy. New York: Arno Press and New York Times.
Vancil, Richard F., Decentralization: Managerial Ambiguity byDesign . New York: Dow Jones-Irwin, 1979.
Weick, Karl E. , "Educational Organizations as Loosely CoupledSystems," Administrative Science Quarterly . Vol.21 (1976), pp.l-19.
Wernerfelt, Birger, "A Resource-Based View of the Firm,"Strategic Management Journal . Vol.5 (1984), pp. 171-180.
Williamson, Oliver E. , Corporate Control and Business Behavior:An Inquiry into the Effects of Organization Form on EnterpriseBehavior . Englewood Cliffs, N.J.: Prentice Hall, Inc., 1970.
Williamson, Oliver E. , "Managerial Discretion, OrganizationalForm, and the Multi-division Hypothesis," in R. Marris and A.Wood (eds) The Corporate Economy . Cambridge, Mass.: HarvardUniversity Press, 1971.
Williamson, Oliver E. and N. Bhargava, "Assessing and Classify-ing the Internal Structure and Control Apparatus of the ModernCorporation." in K.G. Cowling (ed.) Market Structure andCorporate Behaviour . London: Gray-Mills, 1972.
Williamson, Oliver E. , Markets and Hierarchies: Analysis andAntitrust Implications . New York: Free Press, 1975.
Williamson, Oliver E. , "Emergence of the Visible Hand: Implica-tions for Industrial Organization," in A. D. Chandler and H.Daems (ed.) Managerial Hierarchies . Cambridge, Mass.: HarvardUniversity Press, 1980.
Williamson, Oliver E. , "The Modern Corporation: Origins, Evolu-tion, Attributes," Journal of Economic Literature . Vol.19 (Decem-ber 1981), pp. 1537-1567.
Williamson, Oliver E. , The Economic Institutions of Capitalism:Firms. Markets. Relational Contracting . New York: The Free Press,1985.
Zannetos, Zenon, S., "On the Theory of Divisional Structures:Some Aspects of Control and Decision Making," ManagementScience . Vol. 12(4) (December 1965), pp. 49-68.
36